Ollala AI OCR

Local AI-BASED OCR

Ollala AI OCR

A small open-source tool for converting PDFs and images into Markdown with a local Ollama glm-ocr model.

Setup Use Ollala with the web UI

Local processing

PDF and image OCR without a cloud service.

Ollala bridges the PDF gap in Ollama by rendering documents into images first, then sending each page to glm-ocr for Markdown extraction.

  • 01

    PDFs, images, and folders

    Convert a single file, a folder, or a recursive folder tree while preserving multi-page PDFs as one Markdown file.

  • 02

    Quality profiles for scanned documents

    Safe, Balanced, Default Detail, and Higher Detail profiles tune DPI, image size, context, retries, and timeouts.

  • 03

    Web interface and CLI

    Use the web interface for drag-and-drop jobs, or run ocr_to_md.py directly for scripts and batch work.

How it works

A small local pipeline from source files to Markdown.

1

Drop files

Choose PDFs, images, or folders in the local web interface.

2

Render pages

Poppler converts PDF pages into images without loading the whole file into memory.

3

Run OCR

Each page goes to the local Ollama glm-ocr model with live progress logs.

4

Save Markdown

Download the result or write Markdown files to a folder you choose.

Setup

Install the dependencies, then run the local app.

Currently supported on macOS only. You need Python 3, Ollama, glm-ocr, Poppler, and the Python packages from this repo.